The functions of the syntactically non-integrated ‘now’ in Icelandic conversation

Helga Hilmisdóttir, Department of Finnish, Finno-Ugrian and Scandinavian Studies, University of Helsinki

1. Introduction

1.1 Background: and núna in Icelandic conversation

Contemporary Icelandic has two words which can be translated with the English adverb now: and núna. On the semantic level, both types index that the propositional content of an utterance is true during a period which overlaps with the speaking time (cf. Schiffrin 1987: 228 on now in English). Yet, in naturally occurring conversation, and núna tend to occur in different syntactic positions and serve different functions (Hilmisdóttir 2007: 96–148, Hilmisdóttir 2010). While núna typically occurs at the end of turns and refers to a stretch of time (a), tends to occur before the finite verb and indexes a temporal comparison (b), a transition (c), or an emphasis (d). The following four examples show instances found in naturally occurring conversations in a simplified form:

  • Ertu með það opið núna? ‘Do you have it open now?’
  • Hann var of víður. Nú er hann passlegur. ‘It was too big. Now it fits.’
  • Nú er ég búin að fá nóg. ‘Now I’ve had enough.’
  • Nú kem ég við hjá þér á eftir og kref þig um diskinn! ‘Now I will come by later and demand the CD!’

In addition to the temporal meaning, has also developed non-temporal functions as a discourse particle. [1] In this study, discourse particle is used as an umbrella term for words or phrases which, besides their referential meaning, can in some contexts have text-organizing or interpersonal functions (cf. e.g., Foolen 1996). The items which function as discourse particles derive from different lexical categories, and they form a very diverse group of words. Also, items which can function as discourse particles have at times ambiguous meanings or functions, and it is not always clear whether they should be interpreted as lexical words or as particles.

Discourse particles differ considerably in regard to function and syntactic placement (e.g., Brinton 1996: 32–38). In previous research (Hilmisdóttir 2007, Hilmisdóttir 2011), I argue that the discourse particle should be divided into three subcategories based mainly on syntactic criteria. Using Auer’s interactional adaptation (1996: 296) to Diderichsen’s (1962 [1946]) model of syntactic fields, I parse each turn into five different “syntactic fields”. [2] The core syntax contains a “front field” which is the slot before the finite verb. The “middle field” contains the finite verb and possibly a subject and an adverbial. The “end field” contains an object or predicate and possibly an adverbial. The slots surrounding the core syntax are referred to as the “pre-front field” and “post-end field”, respectively, and these two fields are typical slots for interactionally sensitive items such as discourse particles (Auer 1996, Hilmisdóttir 2007: 230–266).

Table 1 shows the possible syntactic positions of the discourse particle , the darker shade marking the core syntax. These examples do not derive from the data. The table does not include examples where is used as a turn of its own.

Table 1. The syntactic position of .

Core syntax

Pre-front field

Front field

Middle field

End field

Post-end field




be.3 PRT

is [NÚ]

frá Íslandi
from Iceland.DAT

from Iceland




er hann
be.3 he

he is

frá Íslandi
from Iceland.DAT

from Iceland








frá Íslandi
from Iceland.DAT

from Iceland

The first two examples in Table 1 are integrated into the core syntax, in the front and middle fields respectively. These particles function as “tone particles” (Hilmisdóttir 2011). The function of these instances is to give the utterance in which they are employed a tone of determination. [3] Notice that in the second example, in which occurs in the front field, the subject of the utterance (hann ‘he’) is shifted to the middle field due to the Icelandic verb-second rule (Thráinsson 2007). In other words, the tone particle is syntactically integrated and fills the front field slot in the same way as a nominal phrase or an adverbial phrase would.

The third example occurs in the pre-front field, and it functions as an “utterance particle”. Such particles are not syntactically integrated in the core syntax. As Table 1 shows, the employment of in the pre-front field does not cause a shift in the word order. Nonetheless, is attached to the utterance by means of projection, that is, it strongly indicates that the turn is still in progress. in the pre-front field projects one of two things: (i) a continuation of a turn produced by the same speaker or (ii) a response to the interlocutor’s turn (Hilmisdóttir 2007: 230–268). [4]

Lastly, instances of which are employed as a turn of their own are referred to as “response particles” (not shown in Table 1) (cf. Hilmisdóttir 2007: 269–323 on dialogue particles). These cases function as pragmatically complete turns by themselves and function in some ways like the English discourse particle oh (cf. Heritage 1984, Heritage 1996, Heritage 2002, Local 1996).

1.2 Aims and disposition of this study

The aim of the present study is to describe the functions of when it is employed as an utterance particle and response particle in talk-in-interaction. The study is synchronic in nature and focuses on the use of in contemporary Icelandic. However, after a sequential analysis of the data, I will also discuss the possible connection between the different types of , and suggest that the particle has emerged diachronically from a temporal adverb to become an eliciting response particle. This aim is reflected in the organization of the study.

The article is organized as follows: Section 2 briefly sketches the methods and data used in the study. Section 3 discusses the theoretical concepts. Section 4 presents the results of my analysis, beginning with instances of as an utterance particle and followed by as a response particle. Section 5 discusses whether the analysis of can tell us something about a possible grammaticalization process in the past. In section 6, the article ends with concluding remarks.

2. Methods and data

The methodological framework of this study is interactional linguistics (e.g., Selting & Couper-Kuhlen 2001, Hakulinen & Selting 2005, Lindström 2006). Research within interactional linguistics uses methods and theoretical concepts which were first developed within conversation analysis (=CA) (e.g., Schegloff 2007). Hence, while a CA-study typically takes social action (such as repairs and apologies) as a starting point, a study within interactional linguistics focuses on particular linguistic features (such as a specific word or word order). As all studies within interactional linguistics, this research is based entirely on empirical evidence. By using excerpts from authentic conversational data, the use of the non-temporal is described from an interactional perspective. The data on which the study is based consists of 14 hours and 11 minutes of recorded conversation. (Audio files for the private conversations have not been included in this publication due to privacy issues.) Table 2 shows information about the recordings included in the study.

Table 2. The database.

Corpus Form of interaction Situation Participants Duration
Friends (1996) Face-to-face Evening gathering (audio recording) Four young women in their mid or late twenties 146 min.
Reunion (1998) Face-to-face Evening gathering (audio recording) Six young women in their early twenties 90 min.
PTC (2003) Telephone calls Private telephone conversations (audio recording) One woman in her early thirties and various other participants 78 min.
ITC (2003) Telephone calls Institutional telephone conversations (audio recording) One woman in her early thirties calling different institutions 17 min.
Teens (1996, 1998) Face-to-face conversations, telephone calls Radio show for teenagers (audio recording) Several young adults 100 min.
The Soul of the Nation (1998) Telephone calls Phone-in program on radio (audio recording) Moderator, studio guests and various callers phoning the program 310 min.
Elections (1996) Face-to-face Political television debate (audio-visual recording) Five Candidates for the presidency, six invited reporters, and one moderator 110 min.
Total duration approximately 14 h. 11 min.

As Table 2 shows, the conversations in the database represent a wide range of conversational settings, including leisure time conversations, a collection of private and institutional telephone calls, three radio shows led by teenagers, ten episodes of a phone-in program on radio, and, finally, a political debate on television. The conversations are all authentic and they were not set up for the purpose of research. The transcription is based on the transcription conventions developed in Hilmisdóttir (2007), and the conversations were transcribed by the author.

3. Central concepts: units of conversation

In this study, concepts that have to do with the construction of turns play a central role. The smallest units of conversation are referred to as “turn-constructional units” (=TCUs). A TCU may vary in length; it can consist of a single lexical item, a phrase, a clause, or a syntactically organized utterance (cf. Sacks, Schegloff & Jefferson 1974: 702). After each TCU there is a “transitional relevance point” (TRP) in the conversation, and at every TRP the discourse floor is open and a speaker-change may occur. If other participants do not claim the floor, the same speaker may continue talking.

A turn consists of at least one TCU. However, as I will show in the analytic part of this study, defining a turn is not always an easy task. A working definition for a turn has been provided by Linell (1998: 159). According to him, a turn is ‘a continuous period when one speaker holds the floor, and the corresponding dialogue contribution is then those verbal and non-verbal actions taken by him during this period, designed to be part of the jointly attended discourse floor, and/or taken up as significant contributions to this floor’. This definition, however, is in many ways problematic, especially in regard to turns which are expanded upon after a possible completion. In this study, such expansions are referred to as “continuations” (see section 4.1.1).

4. Analysis of as an utterance and response particle

There are 557 instances of temporal and non-temporal in the database. Table 3 shows the distribution between the different functions (for more detailed information on the distribution of , see Hilmisdóttir 2007: 325).

Table 3. The functions of in the database.


Temporal adverb
(syntactically integrated)

Tone particle
(syntactically integrated)

Utterance particle
(in the pre-front field)

Response particle
(as a turn of its own)

In total

















































As seen in Table 3, 32 instances of function as utterance particles and 73 instances function as response particles. These two categories are small compared to the syntactically integrated instances, which occur 656 times in the database. However, despite the relatively small number, these instances reveal certain patterns in the use of as an utterance particle and response particle. In the following sections, I will show instances typical of each category, and discuss their sequential placement and prosodic realisation. I will begin my analysis by looking at as an utterance particle.

4.1 as an utterance particle

In the database, there are 32 instances of which occur in the pre-front field and function as utterance particles. These instances can be divided into two main categories: (a) “continuing ” which projects a TCU or turn which ties to the same speaker’s prior turn, and (b) “responding ” which projects a TCU or a turn which ties to the interlocutor’s prior turn. Table 4 shows how the two categories are distributed in the database.

Table 4. The distribution of continuing and responding in the pre-front field.


























11 [5]











As Table 4 shows, the continuing seems to occur only in the moderated media conversations. In such conversations, the turn allocation is generally controlled by one person, and the turns tend to be longer than in normal everyday conversations, in which there is more competition for the floor and turns tend to be shorter (Sacks, Schegloff & Jefferson 1974: 701). This can explain the uneven distribution of the continuing . In contrast, the responsive is distributed rather evenly in the data, and it does not show any clear tendency to occur in a particular data type.

4.1.1 Continuing

I will begin my analysis of as an utterance particle by focusing on the instances that occur in continuations. In (1), a woman has called the program and expressed her wish to discuss a political debate between five presidential candidates, which was broadcast on television the night before. When the excerpt begins, the caller is assessing the first candidate (line 1). The key for transcription and glossing can be found in appendices A and B, respectively. Instances of are left untranslated but marked with boldface and with an arrow in the margin:

(1) Candidates that talk too much (Soul 11.06.96)
(M = Moderator; R = Ragnhildur, the caller)

(Listen to Example 1 (ex1.mp3))

01 R Sigurður Helgi með sama raulið eins og vanalega?
1nameM 1nameM with same song.DEF as and ususal
Sigurður Helgi with the same song as usual

02 (0.5)

03 R og hann var alltaf öðru (.) hverju kon- eh kominn á Alþingi¿
and he be.3.PT always once in a while com- eh come.PP on Parliament
and he was all the (.) time acting as if he was a member of the parliament

04 (1.0)

05 R og ha- maður er nú búinn að hlusta nóg á hann¿
and h- man be.3 NÚ finish.PP to listen enough on he.ACC
and h- I have heard him enough

06 (0.8)

07 R svo mann langar ekkert til þess að halda því áfram¿
so man want.3 nothing to it.GEN to hold it.DAT forward
so I don’t want to continue that

08 (0.8)

→ 09 R Nú:#e::# (0.2).h /Svo er þa- o svo er hún Sigrún¿
.h then be.3 i- and then be.3 she 1nameF
.h then it was Sigrún

10 (0.3)

11 M m

12 (0.6)

13 R og hún var nú bara með mesta: lengsta: #þa:::# (.)
and she be.3.PT NÚ just with most longest it-
and she had NÚ the longest, i-

14 talþáttinn¿
speaking time.DEF
speaking time

15 (0.7)

16 R að það var svoleiðis að það var alveg sama hver talaði þá
that it be.3.PT such that it be.3.PT completely equal who speak.3.PT then
and it was such that it didn’t matter who spoke

17 tók hún alltaf frammí og sag- og sagð#i:# og bað að lofa
take.3.PT she always forward+in and sa- and say.3.PT and ask.3.PT to allow
she always cut them off and sa- and said and asked them to allow

18 sér að komast að? .hh mér finnst það nú ekki nein kurteisi
her.REF.DAT to come.MV VPrt .hh I.DAT find.MV it NÚ not any politeness
her to get the floor, I don’t think it is polite

19 þessum (fréttamanni) sem stjórnar vera að láta þetta vera dálítið
this.DAT (news reporter) who control.3 be to let this be a bit
of this (news reporter) who is in charge to let this be a bit

20 jafnt?

21 (1.2)

→ 22 R .h /Bergsveinn var nú sold#i:ð# svona sann/gjarn o (.) .hh en
.h 1nameM be.3.PT NÚ a bit so fair and .hh but
.h Bergsveinn was NÚ a bit like fair, and .hh but

23 svo Jóhann hann hann va- talaði nú voða stutt í einu?
then 1nameM he he b- speak.3.PT NÚ very short in one
then Jóhann, he he wa- spoke NÚ very short each time

Throughout excerpt (1), the caller experiences an apparent lack of response from the moderator. Instead of having a dialogue with an exchange of turns, the caller finds herself in a situation in which the co-participant does not show any attempt to take the floor or to give any kind of feedback.

After the first negative statement about the first candidate, there is a 0.5 second pause (line 2). During this pause, the moderator shows no intention of taking the floor or providing any kind of feedback, so the turn is redirected to the caller (line 3). The caller then accounts for her negative assessment, but a response from the moderator continues to be absent. Notice how the caller produces four different TCUs, each ending with rising or half-rising intonation and followed by considerably long pauses (lines 1, 3, 5 and 7). She maintains the continuity between the different parts of her argument by using devices such as the connectives og ‘and’ and svo ‘then’. The prosody and the pauses suggest an orientation towards the co-participant. [6] This suggests a lack of agreement between the two interlocutors which can be explained by the institutional role of the moderator as a neutral commentator.

When the first subtopic is exhausted (lines 1–7), there is a 0.8 second pause (line 8). Again, the moderator remains silent and the discourse floor is redirected to the caller. This time, the caller initiates her turn with the particle nú. The particle is produced with a prolonged vowel, creaky voice, a rising-falling intonation contour, and it is delivered as a separate prosodic unit. Furthermore, it is immediately followed by creaky glottal sounds (for information on the prosody of initiating continuations, see Hilmisdóttir 2007: 247). In this instance, projects a continuation of the preceding turn; it has a floor-holding function, and at the same time it introduces a new topic and constructs continuity. The transition between subtopics is further emphasized by the use of the connector svo ‘then’.

Following the caller’s turn in line 9, the moderator provides the first response in this sequence (line 11). Unlike the previous pauses during which the moderator withheld any response, this slot follows a neutral statement with which the caller is introducing the next topic. Thus, the moderator’s response simply acknowledges the transition between two subtopics and does not indicate any support for the negative assessments the caller is making. In other words, this is a safe slot for an acknowledgement token.

In the subsequent talk, the caller continues to criticise the second candidate without the moderator showing any intention of responding to what the caller has been saying. After two unsuccessful attempts of receiving a response (lines 14 and 20), there is a 1.2 second gap in the conversation. After the gap, the caller introduces the third candidate, and again she signals a move between subtopics by using . This time, is unstressed and carries less prosodic prominence. is short and has a rising intonation. The word that follows, i.e., the name of the third candidate, begins with a clear pitch step-up. Unlike in line 9, there is no audible pause between and the remainder of the turn.

To sum up, there are two instances of the particle in this sequence. Both instances signal a shift between two different subtopics – the different presidential candidates – and hence, the particle contributes to the continuity of the cumulative structure (cf. also Schiffrin 1987: 238 on now in English). As I already pointed out, does not involve any speaker-shifts in this interaction. The co-participant simply passes her opportunity to take the floor, and the floor is redirected to the prior speaker. The turn which contains is a direct continuation of the caller’s prior turn.

As shown in (1), can occur in continuations in which speakers expand their own turns. In this particular case, the co-participant does not give any comment or feedback between the first part of the turn and the continuation. She remains totally silent the whole time. In other cases, the continuing does a “skip-tie”, that is, it ties to a turn which is not immediately prior to the continuation. This is the case in (2). In this excerpt, which is also drawn from the phone-in program The Soul of the Nation, is used to construct continuity between two non-adjacent turns. In this particular episode of the program, listeners have the opportunity to phone in and discuss car problems with a professional mechanic. Here, the caller is telling a story about a woman who bought a defective car from a car dealer:

(2) The new car (Soul 23.05.96)
(M = Mechanic; I = Ingi, the caller)

(Listen to Example 2 (ex2.mp3))

01 I ég er að tala um allar tE:ngingar bæði í varaljósum
I be.3 to talk about all wires both in spare lights.DAT
I am talking about all wires, both in the spare lights

02 og öðru.
and other.DAT
and otherwise

03 (0.3)

04 M jájá þetta er [alveg- ]
PRT this be.3 completely-
yes, this is just-

→ 05 I [ Nú ]svo ég (0.6) tók það nú að mér fyrir
NÚ NÚ so I take.1.PT it NÚ to I.DAT for
NÚNÚ so I said I would take care of this for

06 ana hún er g- ett+er gömul kona?
she.ACC she be.3 o- this+be.3 old woman
her she is o- this is an old woman

((3 turns omitted))

07 I <H A N N D U G Ð I Í N Í U S T Ö R T.>
he last.3.PT in nine starts
it lasted for nine starts

08 (0.3)

09 M noh

10 (0.3)

11 I Ég m-merkti við það á mælaborðinu.
I m-mark.1.PT VPrt that on dashboard.DAT.DEF
I marked it on the dashboard

12 M ja: ég skal segja þér það
PRT I shall.1 tell you.DAT it
I’m so surprised!

→ 13 I Núnú (0.5) Svo kom ÞRIÐJI startarinn
NÚNÚ then come.3.PT third starter.DEF
Núnú then came the third starter

14 (0.2)

15 M já

16 I .hhh svo fóru tveir svissar¿
.hhh then go.3.PT two ignitions
.hhh then two ignitions broke down

Ingi is the primary speaker in this excerpt, and the mechanic adds a few short comments without being allowed to take the floor completely. In this excerpt, the use of núnú can be analyzed in terms of “grounding” (Hopper 1979: 214, Hopper & Thompson 1980: 280). Hence, the excerpt contains elements which are parts of the “foreground” and elements which are parts of the “background”. The foreground is a sequence of events which take place on a time line (Hopper 1979: 214). The foreground elements are preceded by núnú twice in Ingi’s story (lines 5 and 13). The background, in contrast, does not directly contribute to the temporal progress of the story. Instead, it provides assessments or comments on it (Hopper & Thompson 1980: 280).

Ingi’s first turn (lines 1 and 2) is part of the story’s background, and it provides more detailed information on the condition of the car. At this point, the turn seems to have reached a completion and subsequently the discourse floor is opening up. Yet, when the mechanic starts taking the floor (line 4), producing an assessment which seems to be designed to claim strong affiliation with the speaker, the caller starts an overlap, and the mechanic yields the turn (line 5). Ingi’s overlap begins with the utterance particle núnú (line 5). [7]

The particle núnú in line 5 is produced as a separate intonation unit with rising intonation, uttered in a determined tone of voice. It is directly followed by the consecutive conjunction svo ‘so’ and a syntactically organized utterance. By employing núnú, Ingi links his turn in line 5 to his turn in line 2 without acknowledging the mechanic’s attempts to take the floor. The particle signals a progress in the linear story and marks a transition from the background to the foreground of the story. Thus, although the co-participant has made an attempt to contribute to the conversation, núnú clearly marks a continuation of the speaker’s own turn: it performs a skip-tie.

Ingi then continues his story and tells the mechanic about his problems (turns omitted). He starts to show his irritation by speaking in a louder voice (line 7). The mechanic responds with an exclamatory utterance which functions rather as a backchannel than a full-fledged turn (line 12). Ingi then takes the floor again and continues his story without acknowledging the mechanic’s contribution (line 13). The first prosodic unit in this continuation consists only of the particle núnú produced with a rising-falling contour. As in the previous instance, núnú projects a move forward in the story, and at the same time it ties the continuation to the same speaker’s last utterance.

In (2), the caller is telling a story, and he uses the utterance particle núnú twice. In both instances, núnú is employed strategically to index a progression in the story and to give the recipient a cue that the different parts of the story belong to one coherent narrative. The use of núnú directs the focus towards the speaker’s own turn, and it links the turn or the turn-continuation he is producing to his prior turn. To be more precise, núnú occurs in slots where the speaker is turning back to the temporal narrative after short digressions to background information. In both instances, núnú is preceded by a side-comment which is not a part of the temporal development in the story.

So far, I have looked at excerpts in which the particle has occurred twice, both times signalling progress in a cumulative structure. In both excerpts, both instances of occur in environments with clear similarities and show shifts between two different parallel parts. In (1), there are shifts between two different subtopics which are equally important to the argument, and, in (2), there are shifts between the background and the foreground in a story. In discussing excerpt (3), I will look at a different usage of . This usage does not show parallel parts but, rather, displays attention to the main points in a larger project.

Consider excerpt (3), again drawn from a conversation from a phone-in program. When the excerpt begins, the caller has just finished explaining that his wife has lost her glasses:

(3) Lost glasses (Soul 03.06.96)
(M = Moderator; B = Birgir, a caller)

(Listen to Example 3 (ex3.mp3))

01 B .hh O:g annaðhvort hafa þau dottið þar ú:r vasa hennar
.hh and either have.3.PL they fall.PP there from pocket.DAT she.GEN
.hh and they have either fallen out of her pocket

02 á planinu?
on parking lot.DAT.DEF
in the parking lot

03 (0.4)

04 M um[hum]

05 B [E]ða þá þegar hún fór að heimsækja systir sína í Lönguhlíðinni?
or then when she go.3.PT to visit sister she.REF in Langahlíð.DAT.DEF
or when she went to visit her sister on Langahlíð

06 M mm

07 (0.5)

08 B hana ##- ana Sóleyju #í# Lönguhlíð ellefu?
she.ACC ##- (s)he.ACC 1nameF in Langahlíð eleven
Sóley in Langahlið eleven

09 M [::]

→ 10 B [.hhh]h #::# #e:#f að einhver hhhh hefur (0.3) glapist á að
.hhhh #nú::# if that someone hhhh have.3 confuse.MV on to
.hhhh #nú::# if someone hhhh has accidentally

11 taka upp þessi gleraugu að þá væri gott að .hhhh ## fá
take VPrt these.ACC glasses.ACC that then be.3.SUBJ.PT good to .hhhh #to# get
picked up glasses it would be good to .hhhh to have

12 þau inn í afgreiðslu hh sundlaugarinnar í Breiðholtinu?
they.ACC in in service desk.ACC hh swimmingpool.GEN.DEF in Breiðholt.DAT.DEF
them taken to the service desk hh in the pool in Breiðholt

Just as in excerpts (1) and (2), excerpt (3) contains one speaker holding the floor while the co-participant provides feedback. However, instead of signalling continuity between different subtopics or between equally important parts of a story, this instance of marks a transition between different phases.

At the beginning of (3), Birgir explains where and when the wife may have lost her glasses. All the TCUs have rising terminal contours (lines 1, 2, 5 and 8), and they all generate feedback from the moderator (lines 4, 6 and 9). Then Birgir moves on to the next phase in his story. He marks the shift with an audible in-breath followed by an instance of in the pre-front field. is produced with level intonation, a prolonged vowel, and creaky voice (line 10).

The particle in (3) initiates a TCU which shifts the focus from background information to the main point of the story: what people should do if they find Birgir’s wife’s glasses. The change is further highlighted by means of a verb tense shift from past tense (lines 1–8) to perfect (line 10) (cf. Wide 2002). By using the particle , Birgir indexes a move forward in the story at the same time as he presents the upcoming turn as a continuation. Again, ties the turn in progress to the prior one and indicates a relationship between the two parts. By using in this slot, the speaker emphasizes the upcoming TCU and indicates that it contains important information.

So far, I have discussed how as an utterance particle may be used as a device to construct continuity between two turns or a turn and its continuation. As I have mentioned, such instances tend to ignore possible contributions by the co-participant, and they tend to focus on the speaker’s own production (cf. Schiffrin 1987: 245 on the egocentric nature of the English now). In the following section, I will show examples of which are more responsive in character since they occur after speaker-shifts.

4.1.2 Responsive

Around half of the instances of in the pre-front field are employed immediately after a speaker-shift has taken place. Instead of tying the new turn to a speaker’s own turn, these instances typically initiate a response to the co-participant’s prior turn. Consider the following excerpt drawn from an everyday conversation between three friends. In (4), Sunna is talking about her graduation party which she plans to have immediately after her formal graduation ceremony. Nanna estimates that Sunna will be home at five o’clock (line 1):

(4) Planning the graduation (Friends)
(N = Nanna; S = Sunna; one silent participant)

01 N °.[þú ert° komin heim fimm¿
PRT you be.2 come.PP home five
yeah you are at home at five

02 S [mt

03 S ## nei >við erum að fara að taka m-< (0.3) ég fer í
PRT PRT we be.3.PL to go to take ph- I go.1 in
well no, we are going to take ph- I will go to

04 myndatöku klukkan fimm¿
photography.ACC clock.DEF five
the photographer at five o‘clock

05 (0.2)

→ 06 N Nú: varstu ekki í myndatöku á fö[studaginn.
were.2+you not in photography.DAT on Fri[day.ACC.DEF
didn’t you go to the photographer on Friday?

07 S [Neinei #semst#
no, it‘s

08 hópmyndatöku af [(.) s+st læknahópnum
group picture taking.ACC of PRT medical group.DAT.DEF
a group picture of y‘know the medical students

09 N [já

10 N °.::°

Sunna corrects Nanna by pointing out that the photographs will be taken at five o’clock (lines 3, 4). Earlier, Sunna told the other women about her family’s visit to the photographer, and, therefore, this new piece of information may sound peculiar.

Thus, by using the particle , Nanna indexes that she has received information which contradicts her assumptions. She has received new information (Sunna is going to the photographer) which does not match her previously acquired knowledge (Sunna has already been to the photographer). In order to resolve the problem, Nanna initiates a repair sequence which starts with an instance of nú. Nú is delivered with a stretched vowel and a rising intonation contour, which projects a continuation that identifies the problem source (line 6).

The responsive does not only project questions and repair sequences. In some cases, the particle is employed before answers to questions. Consider excerpt (5) drawn from a telephone conversation between two friends. Here, Arnar asks Erna jokingly whether she would like to go sunbathing with him. Erna responds by asking him where he is located (line 1):

(5) Let’s go sunbathing! (PTC 03.07.03)
(A = Arnar, the caller; E = Erna, the called)

01 E mt. Hvar ert þú eiginlega,
mt. where be.2 you actually
where are you actually

02 (0.6)

03 A #É:g,#

04 E já.

05 (0.8)

→ 06 A Nú’ #á#Ingholtunum¿
on Thingholt.DAT.DEF
in Thingholt

07 (1.3)

08 E Þa e sko ekkert gEggjað veður í Vesturbænum,
it be.3 PRT nothing crazy weather in Vesturbær.DAT.DEF
the weather in Vesturbær is not fantastic at all

Arnar’s proposal to go sunbathing contradicts Erna’s current knowledge about the weather outside. Therefore, she displays some doubts about where Arnar is located (line 1). Notice the stress on the personal pronoun þú ‘you’. By uttering the pronoun with heavy stress, Erna is probably implying that the weather is not very good where she is located. Arnar treats the question as something that is out of the ordinary. This is indexed by a turn consisting of only the pronoun ég ‘I’ uttered with heavy stress and dynamic pitch (line 3). Following a confirmation from Erna and a long pause (line 5), Arnar finally answers her question by telling her that he is located in the Thingholt neighbourhood (line 6). The answer is preceded by an instance of nú.

In (5), initiates an answer to a question from Erna. However, before the question is answered, Arnar seems to question its relevance. In other words, projects an answer that should be obvious to Erna. The particle shows that Arnar was surprised by Erna’s lack of knowledge. Hence, the particle projects an upcoming answer, an answer which is presented as self-evident by the speaker.

In excerpts (4) and (5), the particle clearly projects a continuation of some sort. The turn is not pragmatically complete until something has been added. In (4), the turn is not complete until the speaker has posed a question which initiates a repair, and, in (5), the turn is not complete until the speaker provides an answer to the lingering question from the co-participant. In other words, is used to embed the upcoming utterance in its context rather than to constitute a complete act on its own. However, in some cases, occurs in the pre-front field in such environments and is prosodically delivered in such a way that it could potentially be a turn of its own. This is the case in excerpt (6), which is drawn from the data The Soul of the Nation. The excerpt shows a beginning of a conversation. In the background, traffic noise can be heard, and this causes the moderator to ask where Karl, the caller, is located (line 1):

(6) Speaking on the cell (Soul 04.06.96)
(M = Moderator; K = Karl, a caller)

(Listen to Example 6 (ex6.mp3))

01 M Hvar ert þú: á fe.
where be.2 you on move.DAT
where are you driving?

02 K .hh Ég er nú eiginlega hvergi á fe¿ Ég er nú kyrrstæður
.hh I be.1 NÚ actually nowhere on move.ACC I be.1 NÚ still
.hh I’m NÚ actually not driving anywhere, I’m standing still

03 eins og er¿
as be.3
as it is

04 (0.2)

→ 05 M Nú: þa er *hehhh eins og þú sért á fleygiferð [í-*
NÚ that be.3 hehhh as you be.2.SUBJ on fast move in-
Nú, it’s as though you are driving at breakneck speed in-

06 K [Ég er í bílsíma
I be.1 in carphone.DAT
I’m on a carphone

07 [að vísu?

08 M [Þú ert í bíl
you be.2 in car.DAT
you’re in a car

Karl responds to the moderator’s question by pointing out that it is based on false premises: he is not driving at the moment (lines 2, 3). The moderator’s response is initiated with an instance of delivered with a stretched vowel and clear rising intonation contour. In the turn that follows, the moderator explains with laughter in her voice the reason for her misunderstanding (line 5). Here, has the function of registering a change-of-state. [8] The speaker has received new information which contradicts her previous assumptions, and, therefore, her previous question has been made irrelevant. The moderator’s knowledge of the situation has changed from uninformed to informed.

In general, the particle , delivered with a dynamic rising or rising-falling prosodic contour, indexes a change of state. In this particular case, the speaker could potentially have stopped after . The particle is pragmatically complete by itself. Yet, in this case, the speaker chooses to explain her reactions. Thus, it is debatable whether this instance should be described as an utterance particle, which embeds the turn it initiates in its context, or whether it should be viewed as a response particle, which does not project any particular type of continuation. In section 4.2, I will discuss instances of which clearly occur as turns of their own.

In this section, I have discussed instances of responsive , that is, instances which are employed after speaker-shifts. In (4) and (5), I show instances which register new or unexpected information. At the same time, these particles cannot occur as turns of their own. Rather, they clearly project a TCU which elaborates on the received information. In (4), the problem source is identified with a question which seeks to clarify a misunderstanding. In (5), the speaker expresses his feeling that the question posed has a self-evident answer. In (6), in contrast, could potentially occur as a turn of its own, although the speaker decides to explain the reason for her surprise. This last example leads us over to the following category of , which consists of instances which are employed as one-word turns.

4.2 as a response particle

In the data, there are 73 instances of which occur as response particles. These instances can be divided into two main categories: (a) instances that register new or unexpected information, or (b) instances that elicit an explanation for partially incomplete or unexpected information. The following excerpt which is drawn from a private telephone conversation between two friends is an example of the first category. Just before the excerpt begins, Erna announces that she is going on a camping trip over the weekend. Arnar responds by asking her where she is going (line 1):

(7) Going camping (PTC 03.08.03)
(A = Arnar, the caller; E = Erna, the called)

01 A °Hvert ertu að fara,°
where be.2+you to go
where are you going?

02 (0.5)

03 E >É er að fara:< í hérna: Skaftafell o: Jökulsárlón
I be.3 to go in PRT Skaftafell.ACC and Jökulsárlón.ACC
I’m going to ehm Skaftafell and Jökulsárlón

04 og þa allt,
and it all
and everything

05 (0.6)

→ 06 A Nú

07 E Já::

08 A Með hverjum ertu að fara,
with who.DAT be.2+you to go
who are you going with?

After a short pause, Erna gives an answer to Arnar’s question (lines 3, 4). Then, Arnar registers Erna’s answer with the particle (line 6). The particle occurs as a turn of its own, and it is uttered with a dynamic rising-falling intonation. As a response to , Erna confirms the information with the acknowledgement token ‘yes’ (line 7). Erna’s acknowledgement token completes this sequence, and Arnar moves on to the next question (line 8).

The instance of in (7) differs significantly from the instances we have seen before. First, it does not project a continuation. The turn begins and is completed with . Second, since the particle occurs as a turn of its own, the prosody is more important for its interpretation. In this case, is produced with a rising-falling intonation. The sequential placement shows us that does not call for any kind of elaboration by Erna. It simply registers that the information provided by Erna has been received by Arnar. Yet, by choosing the particle instead of the unmarked particle ‘yes’, Arnar treats Erna’s information as something that is new to him: the particle indexes a change of state.

Instances of which do not elicit any elaboration or explanation are typically followed by a confirmation (‘yes’) uttered by the provider of the new information. However, in excerpt (8), which is also taken from a private telephone conversation, calls for more than a simple confirmation. The particle has an eliciting function. At the beginning of the excerpt, a new topic is introduced by one of the speakers (line 1):

(8) Unemployed (PTC 11.06.03)
(L = Lárus, the caller; E = Erna, the called)

01 L °.° Ari hann er að verða örvæntingarfullur greyið.
PRT 1nameM he be.3 to become desperate poor thing.DEF
Ari is becoming desperate the poor thing

→ 02 E Nú:

03 (1.4)

04 L Hann fær enga vinnu.
he get.3 no work.ACC
he can’t find a job

05 E Nei

Lárus produces an utterance which is pragmatically incomplete in its context. He informs Erna that Ari is getting desperate but does not explain why (line 1). As a response to this incomplete statement, Erna produces an instance of (line 2). The particle occurs as a turn of its own, and it is produced with stretched vowel and rising intonation (cf. Hilmisdóttir 2007: 302–311) on with rising intonation). It is followed by a long pause in the conversation (line 3), and, thus, clearly calls for some kind of elaboration or explanation. [9] Despite the long pause following , the producer of the particle waits for an elaboration. It is only after an elaboration has been provided that the producer of closes the sequence.

In section 4.2, I have discussed two instances of which are employed as response particles. In (7), functions as an acknowledgement token which registers new and/or unexpected information. The particle does not call for any elaboration on the part of the co-participant. In (8), in contrast, has a clear eliciting function. Instead of simply registering information provided by the co-participant, it calls for further elaboration. The difference between these two instances is indexed by prosody. While the responsive particle is produced with rising-falling intonation, the eliciting particle is produced with rising intonation.

5. A possible path of grammaticalization

The development of discourse particles such as is a process which normally takes place over an extended period of time, at least decades and even centuries. Ideally, an investigation into how the Icelandic temporal adverb acquired its present functions would need to use diachronic data (e.g., Lehti-Eklund 1990, Lehti-Eklund 2003; cf. Meurman-Solin, this volume). However, although there is an abundance of instances of in Icelandic manuscripts from medieval times, these documents lack the interactional and prosodic dimensions that are vital for the interpretation of a particle such as .

Obviously, synchronic studies on discourse particles such as the one presented in the present study do not give any insights into diachronic developments. Yet, a detailed synchronic analysis combined with the application of theories of grammaticalization [10] may give us some ideas of how non-lexical functions may have emerged through time.

An important concept in grammaticalization theory is the notion of cline, which is a series of gradual shifts or a continuum (Hopper & Traugott 1993: 6). As many scholars have argued (e.g., Traugott 1982: 256; see also Lehti-Eklund 1990, Lehti-Eklund 2003, Lehti-Eklund, this volume), discourse particles (or discourse markers) tend to follow the pattern propositional meanings > textual functions > interpersonal functions. In the case of , the propositional meaning of the word is to anchor an event or a situation in the present moment (Hilmisdóttir 2010).

In certain environments (see section 4.1.1), can have textual functions such as introducing new topics and constructing continuity in longer arguments or narratives. In these cases, rather than anchoring events on a time line in the external world,indexes temporal progress in the discourse. The step between the propositional meanings to the textual function is not a conceptually large one. The two types of simply work on different temporal levels: the time in the external world and the time within the discourse (cf. also Schiffrin 1987: 245). However, in addition to its temporal function within the discourse, the textual also indexes some kind of emphasis at important junctures in the conversation, for instance between subtopics.

The emphasis at important junctures can perhaps explain the step from the connecting to the responsive (see section 4.1.2). The responsive occurs at the beginning of turns, right after a speaker shift has taken place. In these cases, indexes that something in the prior turn contains new and/or unexpected information or is in some way out of the ordinary. However, these instances typically do not function as pragmatically complete turns by themselves. In most cases, these instances clearly indicate a continuation which casts light on which part of the prior turn was unexpected. The degree of surprise can also be indexed by intonation and vowel length (Hilmisdóttir 2007: 278–282). In other words, the responsive serves an interpersonal function.

The next step in the development of is the step from an utterance particle which embeds a longer turn in its context, to a response particle which functions as a pragmatically complete turn on its own (see section 4.2, example (7)). The difference between these two categories is small, and it entails only the possibility of as a response particle to function as a one-word turn. In cases in which occurs as a turn of its own, the prosodic realization becomes even more dynamic than in cases in which projects a continuation.

The final change in the development of is the step from responding to eliciting response particles (see section 4.2, example (8)). As an eliciting response particle, the main function of is to indicate that something in the interlocutor’s prior turn contains something that requires an explanation. In other words, does the same kind of work as the question ‘Why do you say that?’ does. Whether a responding particle or an eliciting response particle, treats the information carried in the prior utterance as something that was unexpected. However, while the responsive simply registers the new information, the eliciting indicates that the information is not complete and that the co-participant has to elaborate on his or her turn. The difference between the responsive and eliciting is indexed by prosody. While the responsive is typically uttered with a rising-falling contour, eliciting is uttered with a clearly rising contour.

The possible path of grammaticalization from a temporal adverb to an eliciting response particle is summarized in Table 5.

Table 5. A summary of the different stages of .


Stage 1    >

Stage 2    >

Stage 3    >

Stage 4    >

Stage 5

Grammatical category

Temporal adverb

Utterance particle

Utterance particle



or sequential

In the adverbial slot (or non-integrated)

Utterance initial, in the pre-front field

Utterance initial, in the pre-front field

Turn of its own

Turn of its own



Prolonged vowel and creaky voice






Sequential function

Temporal index





Functional level







Indexes the temporal deictic center

Indexes a movement along a timeline

Introduces a response to unexpected information

Registers unexpected information

Registers incomplete information

Examples in the study

(not shown)

(1), (2), (3)

(4), (5), (6)



6. Concluding remarks

In European languages, it seems to be a wide-spread phenomenon that words that refer to the present moment acquire particle functions, for example the English now (Schiffrin 1987), the Norwegian (Andvik 1992, Hasselgård 2006), the Polish teraz (Kryk-Kastovsky 1992, Kryk-Kastovsky 1997), the Swedish nu (Lehti-Eklund 1990, Hakulinen & Saari 1995), and the Finnish nyt (Raevaara 1989, Hakulinen & Saari 1995, Sorjonen 2002). Thus, although the study presented here is based on Icelandic data, the findings are relevant for other languages as well. The analysis shows the kinds of functions the temporal origo can acquire, and how the various functions are connected to one another.


[1] Discourse particles have also been called pragmatic particles (Östman 1981), pragmatic markers (Brinton 1996, Fraser 1996, Andersen 2001, Aijmer & Simon-Vandenberger 2006), discourse markers (Schiffrin 1987), and cue phrases (Hirschberg & Litman 1987), just to name a few.

[2] I have adapted Auer’s model to modern Icelandic syntax. See also Thráinsson (2007: 19) for an application to written Icelandic.

[3] These instances of have many similarities with modal particles in German and Swedish (cf. König 1991, Andvik 1992, Aijmer 1996). However, since does not index modality in the strict sense of the word, I prefer to use the term “tone particle”.

[4] These instances of have some similarities with the use of now as a discourse marker in English as described e.g. by Schiffrin (1987: 228–266) and Aijmer (2002: 57–96).

[5] Three of these instances are reduplicated as núnú.

[6] Unfortunately, no systematic studies on the function of intonation in Icelandic conversation are available. My previous studies, however, have shown that rising intonation is used in many different contexts. In some cases it signals an incomplete turn, while in other cases it seems to signal a wish for an agreement or a positive response.

[7] The second instance of in line 5 is an example of a tone particle (Hilmisdóttir 2011).

[8] In excerpt (6), the utterance particle has many similarities with the change-of-state token oh in English (Schiffrin 1987: 228–266, Heritage 1984, Heritage 1996, Heritage 2002, Local 1996).

[9] This function has many similarities with the use of nu in Hebrew (Maschler 2009: 52).

[10] Traugott (1997:15) defines “grammaticalization” as ‘the process whereby lexical material in highly constrained pragmatic and morphosyntactic contexts becomes grammatical, and already grammatical material becomes more grammatical.’ Within studies on discourse particles, some scholars prefer to use the term “pragmaticalization” (cf. e.g., Diewald 2011). Pragmaticalization has been defined as the grammaticalization of discourse functions.

Appendix A: Transcription symbols

(0.5)Silences measured in tenths of seconds
(.)Micro pause, i.e., a pause no longer than 0.2 seconds
[Overlapping talk begins
]Overlapping talk ends
[[Two or more speakers start simultaneously
< >Talk inside is done with a pace slower than surrounding talk
> <Talk inside is done with a pace faster than surrounding talk
:::Lengthening of sound. Each colon indicates that the sound has been lengthened by approximately 0.10 seconds
-Sudden cut-off of a sound without completing a word
´A glottal stop or word finished abruptly but not cut off
Capitalized letters signal a beginning intonation
núUnderlined means stressed syllable
/A slash indicates a high onset or a pitch step-up
\A reverse slash indicates a pitch step-down
An arrow pointing down marks a fall in pitch
An arrow pointing up marks a rise in pitch
An arrow pointing forward signifies a level contour
.A full stop indicates a falling terminal contour
,A comma indicates slightly falling terminal contour
?A question mark indicates a rising terminal contour
¿A reversed question mark indicates a half-rising terminal contour
Uppercase indicates emphasis with louder volume
°  °Talk inside uttered with sotto voce
*  *Talk inside delivered with a laughing voice
$  $Talk inside delivered with a smily voice
#  #Creaky voice
.The word is said with in-breath
hAudible aspiration. One h indicates approximately 0.10 seconds
.hAudible inhalation. One .h indicates approximately 0.10 seconds
.mtClicking sound caused by parting of the lips
(xx)Item or word in doubt by the transcriber
((xxx))Comments by the transcriber

Appendix B: Glossing

If there is an equivalent form in English, it has been used. In such cases, the grammatical category is not glossed. The following forms are not indicated in the glossing: (i) nominative case and forms which are identical with the nominal case; (ii) gender of nouns, adjectives, participles, and pronouns; (iii) plural of nominals; (iv) singular of verbs; (v) infinitive; (vi) present tense; (vii) active voice.

1, 2, 31st, 2nd, 3rd person
1nameF1st name, female
1nameM1st name, male
MVmiddle voice
PPpast participle
PTpast tense
REFreflexive pronoun
VPrtverb particle


Aijmer, Karin. 1996. “Swedish modal particles in a contrastive perspective”. Language Sciences 18(1–2): 205–216.

Aijmer, Karin. 2002. English Discourse Particles: Evidence from a Corpus. Amsterdam: Benjamins.

Aijmer, Karin & Anne-Marie Simon-Vandenberger, eds. 2006. Pragmatic Markers in Contrast. (=Studies in pragmatics, 2.) Amsterdam: Elsevier.

Andersen, Gisle. 2001. Pragmatic Markers and Sociolinguistic Variation: A Relevance-Theoretic Approach to the Language of Adolescents. Amsterdam: Benjamins.

Andvik, Erik E. 1992. A Pragmatic Analysis of Norwegian Modal Particles. Summer Institute in Linguistics and the University of Texas at Arlington Publications in Linguistics.

Auer, Peter. 1996. “The pre-front field in spoken German and its relevance as a grammaticalization position”. Pragmatics 6(3): 295–322.

Brinton, Laurel J. 1996. Pragmatic Markers in English: Grammaticalization and Discourse Functions. (=Topics in English Linguistics, 19.) Berlin: Mouton de Gruyter.

Diderichsen, Paul. 1962 [1946]. Elementær dansk grammatik [Elementary Danish grammar], 3rd edn. Copenhagen: Gyldendal.

Diewald, Gabriele. 2011. “Pragmaticalization (defined) as grammaticalization of discourse functions”. Linguistics 49(2): 365–390.

Foolen, Ad. 1996. “Pragmatic particles”. Handbook of Pragmatics, 2nd installment, ringbinder, ed. by Jef Verschueren, Jan-Ola Östman & Jan Blommaert. Amsterdam: Benjamins.

Fraser, Bruce. 1996. “Pragmatic markers”. Pragmatics 6(2): 167–190.

Hakulinen, Auli & Margret Selting, eds. 2005. Syntax and Lexis in Conversation: Studies on the Use of Linguistic Resources in Talk-in-Interaction. (=Studies in Discourse and Grammar, 17.) Amsterdam: Benjamins.

Hakulinen, Auli & Mirja Saari. 1995. “Temporaalisesta adverbista diskurssipartikkeliksi” [From a temporal adverb to a discourse particle]. Virittäjä (=Journal of the Society for the Study of Finnish, Kotikielen seuran aikakauslehti, 99(4)), 481–500.

Hasselgård, Hilde. 2006. “‘Not now’: on non-correspondence between the cognate adverbs now and . Pragmatic Markers in Contrast (=Studies in Pragmatics, 2), ed. by Karin Aijmer & Anne-Marie Simon-Vandenbergen, 93–111. Amsterdam: Elsevier.

Heritage, John. 1984. “A change-of-state token and aspects of its sequential placement”. Structures of Social Action: Studies in Conversation Analysis, ed. by Maxwell J. Atkinson & John Heritage, 299–345. Cambridge: Cambridge University Press.

Heritage, John. 1996. “Oh-prefaced responses to inquiry”. Language in Society 27: 291–334.

Heritage, John. 2002. “Oh-prefaced responses to assessments: a method of modifying agreement/disagreement”. The Language of Turn and Sequence, ed. by Cecilia E. Ford, Barbara A. Fox & Sandra A. Thompson, 196–256. Oxford: Oxford University Press.

Hilmisdóttir, Helga. 2007. A sequential analysis of and núna in Icelandic conversation. (=Nordica Helsingiensia, 7.) PhD dissertation, Department of Scandinavian Languages and Literature, University of Helsinki.

Hilmisdóttir, Helga. 2010. “The present moment as an interactional resource: The case of and núna in Icelandic conversation”. Nordic Journal of Linguistics 33(3): 269–298.

Hilmisdóttir, Helga. 2011. “Giving a tone of determination: The interactional functions of as a tone particle in Icelandic conversation”. Journal of Pragmatics 43(1): 261–287.

Hirschberg, Julia & Diane Litman. 1987. “Now let’s talk about now: Identifying cue phrases intonationally”. ACL '87 Proceedings of the 25th annual meeting on Association for Computational Linguistics, 163–171.

Hopper, Paul J. 1979. “Aspect and foregrounding in discourse”. Discourse and Syntax. Syntax and Semantics, vol. 12, ed. by Talmy Givón, 213–241. New York: Academic Press.

Hopper, Paul J. & Sandra A. Thompson. 1980. “Transitivity in Grammar and Discourse”. Language 56: 251–299.

Hopper, Paul J. & Elizabeth C. Traugott. 1993. Grammaticalization. Cambridge: Cambridge University Press.

König, Ekkehard. 1991. The Meaning of Focus Particles: A Comparative Perspective. London: Routledge.

Kryk-Kastovsky, Barbara. 1992. “The pragmatics of interjections: the case of Polish no”. Journal of Pragmatics 18(1): 123–137.

Kryk-Kastovsky, Barbara. 1997. “From temporal adverbs to discourse particles: An instance of cross-linguistic grammaticalization?” To Explain the Present: Studies in the Changing English Language in Honour of Matti Rissanen, ed. by Terttu Nevalainen & Leena Kahlas-Tarkka, 319–328. Helsinki: Société Néophilologique.

Lehti-Eklund, Hanna. 1990. Från Adverb till Markör i Text. Studier i semantisk-syntaktisk utveckling i äldre svenska. (=Skrifter Utgivna av Svenska Litteratursällskapet i Finland, 561. Humanistiska avhandlingar, 4.) Helsingfors: Svenska litteratursällskapet i Finland.

Lehti-Eklund, Hanna. 2003. “The grammaticalization of alltså and således, two Swedish conjuncts, revisited”. Cognitive Approaches to Lexical Semantics, ed. by Hubert Cuyckens, René Dirve & John R. Taylor, 123–162. Berlin: Mouton de Gruyter.

Lindström, Jan. 2006. “Interactional Linguistics”. Handbook of Pragmatics, ed. by Jef Verschueren & Jan-Ola Östman. Amsterdam: Benjamins.

Linell, Per. 1998. Approaching Dialogue: Talk, Interaction and Contexts in Dialogical Perspective. (=Impact, Studies in Language and Society, 3.) Amsterdam: Benjamins.

Local, John. 1996. “Conversational phonetics: some aspects of news receipts in everyday talk”. Prosody in Conversation. (=Studies in Interactional Sociolinguistics, 12), ed. by Elizabeth Couper-Kuhlen & Margret Selting, 177–230. Cambridge: Cambridge University Press.

Maschler, Yael. 2009. Metalanguage in interaction: Hebrew discourse markers. Amsterdam: Benjamins.

Östman, Jan-Ola. 1981. You Know: A Discourse-Functional Approach. Amsterdam: Benjamins.

Raevaara, Liisa. 1989. “No – vuoronalkuinen partikkeli”. Kieli 4. Suomalaisen keskustelun keinoja I, ed. by Auli Hakulinen,147–161. Department of Finnish, University of Helsinki.

Sacks, Harvey, Emanuel A. Schegloff, & Gail Jefferson. 1974. “A simplest systematics for the organization of turn-taking for conversation”. Language 50: 696–735.

Schegloff, Emanuel A. 2007. Sequence Organization in Interaction: A Primer in Conversation Analysis. Vol. 1. Cambridge: Cambridge University Press.

Schiffrin, Deborah. 1987. Discourse markers. (=Studies in interactional sociolinguistics, 5.) Cambridge: Cambridge University Press.

Selting, Margret & Elizabeth Couper-Kuhlen. 2001. Studies in Interactional Linguistics. Amsterdam: Benjamins.

Sorjonen, Marja-Leena. 2002. “Recipient activities: the particle no as a go-ahead response in Finnish conversation”. The Language of Turn and Sequence (=Oxford Studies in Sociolinguistics), ed. by Cecilia E. Ford, Barbara A. Fox & Sandra A. Thompson, 165–195. Oxford: Oxford University Press.

Thráinsson, Höskuldur. 2007. The Syntax of Icelandic. (=Cambridge Syntax Guides.) Cambridge: Cambridge University Press.

Traugott, Elizabeth. 1982. “From propositional to textual and expressive meanings: Some semantic-pragmatic aspects of grammaticalization”. Perspectives on Historical Linguistics, ed. by Winfred P. Lehmann & Yakov Malkiel, 245–271. Amsterdam: Benjamins.

Traugott, Elizabeth C. 1997. “The role of the development of discourse markers in a theory of grammaticalization”. Paper presented at ICHL XII, Manchester 1995, version of 11/97.

Wide, Camilla, 2002. Perfect in dialogue: form and functional potential of the vera búinn að + inf. construction in contemporary Icelandic. (=PIC Monographs, 3.) PhD thesis, Department of Scandinavian Languages and Literature, University of Helsinki.